Large deviations principle for the invariant measures of the 2D stochastic Navier-Stokes equations on a torus (Q2363156)

The paper under review establishes a large deviation principle for the two-dimensional stochastic Navier-Stokes equation with small additive noise and periodic boundary conditions. More precisely, for \(\varepsilon >0\) the Navier-Stokes equation \[ \mathrm du(t)+ A u(t)\,\mathrm dt+ B(u(t),u(t))\, \mathrm dt=\sqrt{\varepsilon}\,\mathrm d w(t),\quad u(0)=u_0, \] is considered on a two-dimensional torus, where \(A\) denotes the Stokes operator, \(B\) is a non-linearity and \(w(t)\) is a suitable cylindrical Wiener process. It is well-known that the Markov process generated by the above Navier-Stokes equation has a unique and ergodic invariant measure \(\mu_\varepsilon\). The main result shows that the family of invariant measures \( \{\mu_\varepsilon \}_{\varepsilon >0}\) satisfies a large deviation principle as \(\varepsilon \downarrow 0\) with rate \(\varepsilon\) and action functional equal to the quasi-potential \(U\) associated to the controlled deterministic Navier-Stokes equation, which is also called skeleton equation. For the proof, exponential estimates for the family of probability measures \(\{\mu_\varepsilon \}_{\varepsilon >0}\) are derived.